N7 ATMOSPHERIC SCAVENGING(UIASA-CR-150491)

Final Report (IIT Research Inst.)EXHAUST 85 p BC A05/b? AOI CS01 21R onclas

G3/20 57757

~~Ic­3/205775

' 149

V,,­

https://ntrs.nasa.gov/search.jsp?R=19780007226 2020-03-22T05:17:26+00:00Z

Advanced concepts are being used by I IT Research Institute to solve research, development, and design problems for industry and govern­ment through contract research. Our services encompass virtually all of the physical and biological sciences. Prihciple areas are: chemistry, computer sciences, electronics, engineering mechanics, life sciences, mechanics of materials, medical engineering, metals, and management and social sciences research.

The interdisciplinary approach at IITR I brings the latest technology to bear upon the problem-solving process.

Principle office. 10 West 35th Street Chicago, Illinois 60616

Contract No. NAS-31947 Final Report IITRI Report No. C6365-17

ATMOSPHERIC SCAVENGING EXHAUST

National Aeronautics and Space Administration

George C. Marshall Space Flight Center Marshall Space Flight Center, Alabama 35812

Attention: Dr. J. B. Stephens Technical Monitor

Prepared by

Donald L. Fenton Robert Y. Purcell

ITT Research Institute 10 West 35th Street Chicago, Illinois 60616

December 1977

ATMOSPHERIC SCAVENGING EXHAUST

ABSTRACT

Solid-propellant rocket exhaust was directly utilized to

ascertain ranidrop scavenging rates for hydrogen chloride. The

airborne HCl concentration varied from 0.2 to 10.0 ppm and the

raindrop sizes tested included 0.55mm, 1.1mm, and 3.0mm. Two

chambers were used to conduct the experiments -- a large, rigid­

walled, spherical chamber stored the exhaust con~titutents while

the smaller chamber housing all the experiments was charged as

required with rocket exhaust HCl.

The washout coeffecient for rocket exhaust HCl. scavenging

as determined empirically is:

17 65) M R0. 773 A = (5.12 x 10

­

where MA is the mass concentration of HCl (g/m3 ) and R the rain­

fall intensity (mm/hr). The washout coefficient is noted to dis­

play a slight dependence on the HCI concentration.

Surface uptake experiments demonstrated an Ha1 concentration

dependence for distilled water. Sea water and brackish water

HCl uptake was below the detection limit of the chlorine-ion analy­

sis technique employed. Plant-life HCl uptake experiments were

limited to corn and soybeans. Plant age effectively correlated

the HC1 uptake data. Metallic corrosion was not significant for

single 20 minute exposures to the exhaust HCl under varying rela­

tive humidity.

Characterization of the aluminum oxide particles substantiated

the similarity between the constituents of the small-scale rocket

(227 g) and the full size vehicles. Also, a single aluminum oxide

scavenging experiment was conducted with the 1.lmm droplets and

tentatively suggests that large numbers of submicron particles

were collected.

IIT RESEARCH INSTITUTE

ii

FOREWORD

This document is the final report on IITRI's Project No.

C-6365 entitled "Atmospheric Scavenging Exhaust" and was per­

formed under NASA Contract No. NAS-31947 from the Marshall Space

Flight Center. The report presents and discusses the rocket

exhaust HCI scavenging data and the auxiliary experiments con­

ducted.

The authors enthusiastically acknowledge the numerous suggestions and support from the project monitor, Dr. J. Briscoe

Stephens, at the NASA Marshall Space Flight Center. The rocket

motors used to generate the exhaust cloud were obtained from the

Jet Propulsion Laboratory, where the assistance of Mr. Leon Strand

was especially important. Guidance from Mr. Keith Dumbauld of

Cramer Co. was important in modeling the experimental data.

At IITRI, Dr. E. L. Grove painstakingly carried out the

chlorine-ion analysis and is gratefully acknowledged for this

work. Mr. Harry Nichols and Mr. Vernon Hill of IITRI's Metals

Division designed and conducted the metallic corrosion tests.

Dr. Manfred Ruddat, a botanist at the University of Chicago,

served as a special consultant relating to HCI plant uptake and

growth response.

IIT RESEARCH INSTITUTE

iii

The authors are pleased to present this report to the

Marshall Space Flight Center in contributing to the task of

determining the environmental impact of the Space Shuttle Pro­

gram.

Respectfully submitted,

IIT RESEARCH INSTITUTE

Donald L. Fenton Research Scientist Fine Particles Research

Robert Y. Purcell Associate EnvironmentalEngineer Fine Particles Research

Approved by

/eohn D. Steckham Manager Science Advisor Fine Particles Research

lIT RESEARCH INSTITUTE

iv

TABLE OF CONTENTS

Page No.

I. INTRODUCTION 1

2. EXPERIMENTAL APPARATUS 2

2.1 Aerosol Chambers 2 2.2 Rocket Motor and Exhaust Cloud 9 2.3 Rain and Fog Simulation 10 2.4 Rain Collection 18 2.5 Measurement of Hydrogen Chloride

Concentration 19 2.6 Absorption Chamber 31 2.7 Growth Chamber of Experimental Plants 39 2.8 Metallic Coupons 42

3. EXPERIMENTAL PROCEDURE 44

4. EXPERIMENTAL TEST DATA AND RESULTS 47

4.1 Scavenging Test Results

4.1.1 Rocket Exhaust HCI Scavenging Data 47 4.1.2 Correlation of Scavenging Data 51

4.2 Absorption Tests 54 4.3 Plant Uptake Tests and Growth Response 56 4.4 Metallic Corrosion Tests 60 4.5 Characterization and Scavegning of Rocket

Exhaust Aluminum Oxide Dust 64

5. CONCLUSION 70

References 76

liT RESEARCH INSTITUTE

V

LIST OF FIGURES

Page

1 Experimental Apparatus .......... ......... 3

5 Rocket Motor Mounting within 5.49 m (18 ft)

6 Data Sheet for Experimental Test Rocket

9 Liquid Calibration Response for the IITRI HC1

10 Liquid Calibration Response for the Langley HCI"

18 Rocket Exhaust HC1 Scavenging Data for Multiple

20 Rocket Exhaust HCI Deposition Velocity Variation

22 Net Coupon Weight Changes Versus Length of

2 Support Frame for Teflon Comprising Test Chamber 5 3 Photograph of Completed Experimental Chamber . 6 4 Assembly Detail of Diluter/Transport System . 8

Spherical Chamber .... ............. .I.I.. 11

7 Raindrop Generator ... ................. .. 15 8 Raindrop Receptacle .... ............. .... 20

Detector ...... ................. ..... 24

Detector ............................ .. 25 11 Calculation Results for Droplet and Gaseous

Diffusion of HCI Showing Theoretical Magnitude of Separation ............ ......... . 28

12 Schematic Diagram of Air-Dilution System .. 30 13 Design of Duct for HC1 Absorption Experiments . . 35 14 Schematic Diagram of Equipment Used in Preliminary

HCI Absorption Experiment ..... ...... ... 37 15 Surface Absorption Flux of HC1 to Distilled Water

Versus Mean Gaseous HC1 Flux Above Surface . . 40 16 Plant Growing Chamber .... ........... ... 41 17 Corn and Soybeans (behind corn) Responding to

Conditions Inside Growth Chamber . . ...... 43

Raindrop Sizes ...... ................ ... 49 19 Raindrop Terminal Fall Velocity as Determined by

Gunn and Kinzer ...... ................. 52

with HC1 Concentration ................. 57 21 Growth Response of Soybean Seedlings Two Weeks

After Germination ..... ........... 61

Exposure Time ...... ........ .. ........ 62 23 A1203 Particle Size Data for Second Scavenging

Test ....................... 66 24 A1203 Particle Size Data for Third Scavenging Test 25 A1203 Particle Size Distributions Based on Number

68

for Both Airborne and Collected Rain ....... 69 26 Photomicrographs Showing A1 20 3 Particles Collected

by 1.1 mm Raindrops at 3,OOOX Magnification . . 71 27 Photomicrographs Showing A1 203 Particles Collected

by 1.1 mm Raindrops at 10,000X Magnification . 71

A-1 Pure HC1 Scavenging Data for Multiple Raindrop Sizes ....... ................ ..... 75

UIT RESEARCH INSTITUTE

vi

LIST OF TABLES

Page

1 Raindrop Generation System Characteristics ..... ... 17 2 Scale Factor Calibration for the HCl Detectors 22 3 Composition of Laboratory Sea Water .. 32 4 Preliminary Test Results for Absorption Chamber" 38 5 Materials for Corrosion Test Matrix ...... 45 6 Droplet Scavenging Results for Rocket Exhaust HC 48 7 Rocket Exhaust HCl Absorption Test Results . ... 55 8 Summary of Test Plants HCl Uptake Measurements . 59 9 Measured Weight Changes Associated with Exposed

Metal Coupons ......... ............. . 63

lIT RESEARCH INSTITUTE

vii

ATMOSPHERIC SCAVENGING EXHAUST

1. INTRODUCTION

An assessment of the environmental effects associated with

the ground exhaust cloud formed during the initial phase of a Space Shuttle launch is important because a major constituent

of the exhaust cloud is hydrogen chloride. It is well known

that hydrogen chloride can be potentially toxic depending on

the local circumstances.

Investigations conducted recently at both NASA-Langley and IITRI determined the extent of HCI washout for rain storms typi­

cal of the Cape Kennedy launch area. However, this early work

left a serious gap concerning the scavenging of rocket exhaust

HCI. Pellet (1) at NASA-Langley conducted very careful experi­

ments with pure gaseous HCI and the data obtained supported the

Frossling correlation. Knutson and Fenton (2) at IITRI conducted

experiments utilizing real solid-propellant rocket exhaust but

limited the conditions to relatively high HCI concentrations

and a single droplet size. The goal of this work, then, was to

fill the gap -- experimentally evaluate actual rocket exhaust

HCI scavenging under typically encountered conditions for vary­

ing raindrop sizes. Auxiliary experiments were conducted to

determine absorption rates or "uptake" of HC1I for a range of

materials including plant-life and liquids common to the launch

site. In addition, corrosion studies were performed on a rep­resentative set of metallic coupons.

Modifications in the test apparatus were necessary to facil­itate multiple HC1 scavenging and absorption experiments. These

changes greatly expanded the information obtained. The modifica­

tions included limiting the large,spherical chamber for the sole

purpose of storing the exhaust cloud constituents generated by the test rocket. A second experimental chamber, much smaller in volume, served as the location for all the experimental activity.

lit RESEARCH INSTITUTE

1

After the exhaust cloud has formed and reached the term­

inal altitude, prevailing winds cause the cloud to drift and

local wind shear and atmospheric turbulence cause cloud dis­

persion. The stage of cloud history is predictable by the

NASA/MSFC multi-layer diffision models (3). These models per­

mit taking into account a number of meteorological parameters

and allow stratification of the atmosphere. The multi-layer

diffusion models have a provision for calculating precipita­

tion scavenging of H0 from the ground exhaust cloud. The

required input for HC1 scavenging is the washout coefficient -­

a function of only the rain intensity.

2. EXPERIMENTAL APPARATUS

The thrust of this study was the determination of the H1

scavenging rates from a rocket exhaust cloud at concentrations

typical of the ground exhaust cloud. Small scale rockets ignited

within a large, spherical chamber generated the exhaust cloud,

which, at appropriate times, was transported to a second, small

chamber constructed of Teflon. This second chamber is where all

the experiments were conducted. In this way, the HC concentra­

tion was sufficiently diluted to achieve the necessary low con­

centrations. The remaining experimental equipment included the

rain simulator, fog nozzle, rain collecter, absorption chamber,

and plant growth chamber. The purpose of all this equipment was

to confine the exhaust cloud in order to carry out the meaningful

measurements and characterize the exhaust cloud. Special instrumen­

tation-was used to monitor the HCI concentration within the cloud

and measure the particle size distribution of the A1203 dust.

2.1 Aerosol Chambers

Two chambers were used to perform the scavenging and uptake

tests for HCl. The first chamber (large sphere) was used to

store the exhaust gases after the test rocket was ignited. The

second chamber (Teflon bag) was the location of all the experi­

mental activity except surface absorption. Figure I is a schem­

atic diagram depicting the major components of the experimental

system.

lIT RESEARCH INSTITUTE

2

Vent

Exhaust Cloud Storage Chamber

Plant Exposure to Exhaust Cloud

rRocket Motor Motor\ \

Experimental Test Chamber, Teflon

Rain Simulator, Varying

Drop Size\ -%--Small Mixing

:U

'U M

Tau

i Ai

-4

Transpiration Tubel ,,tHumid Air Introduced . Rain Collection Outside

- f Chamber Surface Corrosion Experiment

Figure 1

EXPERIMENTAL APPARATUS

The spherical chamber was fabricated from welded steel plate

and is 5.49m in diameter. This chamber is rated for 5.44 atmos­

pheres of gauge pressure and has a calculated volume of 86.5m.

The inner surface of the chamber was coated with Plasite #7122,

an epoxy-phenolic material resistant to acids, and an oil-base

enamel. A wash-down spray head was installed at the top of the

chamber to clean the chamber after each rocket ignition. The

rocket motor was secured to the chamber wall by means of a stud

fastened to a modified port. The chamber was also fitted with

an exhaust system that when operated maintained a negative pres­

sure within, permitting evacuation of the contents. Evacuation

was performed at the conclusion of each test.

The second chamber, where the scavenging experiments were per­

formed, has an available rocket exhaust cloud supply over a period

of several hours. This time interval was deduced from HCI con­

centration decay and coagulation rates of the A1203 dust. The

significance of the second "experimental" chamber is that it

afforded convenient control over the exhaust cloud concentrations

tested and permitted multiple tests for one rocket firing.

The material used in the construction of the experimental

chamber was FEP Teflon film* (.0127 cm thick) and modified on

one side to accept special adhesives. The experimental chamber

was essentially a bag, and therefore required external support.

Figure 2 gives the outside dimensions of the frame used to

support the bag. The bag was made to contour the basic recti­

linear shape of the support frame by glueing the Teflon sheets

in a way to locate the unmodified surface on the inside of the

bag. A removable door, 61 cm square, was provided to change experiments within the experimental chamber for any one rocket

firing. The door was designed utilizing a special rubber gasket

to provide a tight seal with the remainder of the test chamber.

A plywood floor was located under the lower surface of the bag

in order to support test equipment within the experimental cham­

ber. Figure 3 shows a photograph of the completed experimental

chamber. *Cadillac Plastic and Chemical Co., Dallas, Texas

It RESEARCH INSTITUTE

4

2" x 2" Wood Framing

JI

152.5 /

/ __ ___ _

76 ' / 61

// //

61-2

Figure 2

SUPPORT FRAME FOR TEFLON COMPRISING TEST CHAMBER

Note: All dimensions in cm."

liT RESEARCH INSTITUTE­

5

ORIGINAL PAGE 16

OF POOR QUALITY

Figure 3

PHOTOGRAPH OF COMPLETED EXPERIMENTAL CHAMBER

liT RESEARCH INSTITUTE

6

One problem arose with glueing certain sections 6f the

experimental chamber. The Teflon film was modified on one

side to accept and hold a glueing compound while the opposite

side would not. An etching kit was used to modify the untreated

side so that glueing could also be achieved, thus enabling the

construction of the experimental chamber with a door and still

maintaining a complete enclosure of Teflon. The etching tech­

nique did not function properly. A series of other adhesives

were attempted and one, Fluoro-Plastic adhesive #30, was found

satisfactory. This adhesive is a pressure sensitive contact

glue for binding Teflon and other plastics.

The successful use of the experimental chamber necessitates

the efficient transport of the rocket exhaust from the spherical

chamber to the experimental chamber. A specially designed trans­

port tube was installed to minimize wall losses of both HC1 and

A1203 . The concept used in the transport incorporates transpir­

ation air introduced along the length of the tube. This concept

has been used by Ranade (4) with good success to transport aero­

sols in other applications. The transpiration air sheath, which

permits the efficient transport of the aerosol, was generated by

supplying air via a manifold (4 parts along total 46 cm. length)

along a porous tube. Figure 4 is a diagram showing the assembly

details of the dilution/transport system.

To accomplish the transport of the exhaust from the spheri­

cal chamber to the experimental chamber, the negative pressure

generated by laboratory's exhaust fans was utilized. PVC pipe

and valves were used to transport and regulate the rocket exhaust

to the experimental chamber. The expelled exhaust was transported

via a flexible 5 cm. hose to the laboratory's air filtration

box which provides the source of vacuum. Special internal bag

supports coated with epoxy-resin paint and glued wood slats pre­

vented the bag from dollapsing upon discharge.

A fan was installed inside the experimental chamber to pro­

vide good mixing of the bag's contents. The fan itself was fab­

ricated from sheet steel and bolted to a steel shaft. Both the

lIT RESEARCH INSTITUTE

7

A - Porous tube spacer

B - Porous tube

C - Outer shell manifold

©C

N

brass collar

-4 n/i

K

03

z

C

-.4i

-1

Cn't

Figure 4

ASSEMBLY DETAIL OF DILUTER/TRANSPORT SYSTEM

fan blades and shaft were coated with epoxy-resin paint to mini­

mize HCI absorption and consequent corrosion. The Teflon bag

provided a seal as the hole was slightly smaller than the shaft

diameter. The driving motor had a variable speed control thus

providing numerous levels of agitation. The fan blade length

was approximately 10 cm and the shaft length roughly 40 cm. With

the introduction of fog, the mixing action was observed to vary

from mild to vigorous.

2.2 Rocket Motor and Exhaust Cloud

The rocket motors used to generate the exhaust cloud within

the spherical chamber were obtained from the Jet Propulsion Labor­

atory in Pasadena, California.* Each rocket motor contained

approximately 0.277 kg bonded-in propellant grain, with a 5.19 cm

diameter axial perforation. The propellant outside diameter was

7.6 cm while the length was 5.16 cm. The propellant composition

was as follows (5):

Ammonium perchlorateAluminum

70% by weight 16%

PBAN 14%

Minor constituents of the propellant grain were changed from the

previous scavenging experiments (2). One modification included

the addition of about 0.1 to 0.2 percent iron oxide to regulate

the burning rate of the rocket motor. In this study, iron oxide

was shown, through particulate sizing of the exhaust cloud, as

not influencing the overall particle loading or particle size dis­

tribution. Other trace constituents were also modified; but,,

because of their low concentrations, they are not considered sig­

nificant for scavenging experiments,

The rocket exhaust gases discharged through a converging­

diverging nozzle with an expansion ratio of 4:1 and a throat

diameter of 0.927 cm. A spent rocket mounted within the cham­

*Arrangements for the delivery of these rockets to IITRI were made by Dr. J. Briscoe Stephens and Mr. H.R. Hope of NASA/George C. Marshall Space Flight Center and Mr. Leon Strand of the Jet Pro­pulsion Laboratory. Pertinent JPL drawings are D9041893 and D9041612.

[IT RESEA-RCH INSTITUTE

9

ter is shown in Figure 5. A typical trace of the rocket motor

combustion chamber pressure is displayed in Figure 6. During the

test rocket burn, the chamber pressure rise was sufficiently rapid

to permit the generation of typical exhaust constituents. Among

all the pressure traces received, burn time variations were observed

on the order of 10-15%.

As the spherical test chamber contained approximately 102 kg

of air prior to rocket ignition and the rocket 0.227 kg of pro­

pellant, the dilution factor within the spherical chamber was

approximately 450 on a mass basis. Assuming that all the chlor­

ine goes to HCl and all the aluminum goes to Al203, the follow­

ing initial'concentrations in the spherical chamber were expected

(not measured):

HU1: 49.1g, 0.567 g/m3 , 386 ppm (volume basis)* A12 03 : 68.7g, 0.794 g/m3

In the above calculated concentrations, after-burning was consid­

ered to be complete.

2.3 Rain and Fog Simulation

The method of raindrop generation used produces uniformly

sized drops. This enables the determination of the HC1 scav­

enging rate for a specific raindrop size. With the HC1 scav­

enging rate varying with droplet size, the effective scavenging

rate for a particular rainstorm can be found by knowing the

raindrop size distribution. This method of approach is taken

because it is not feasible in the laboratory to duplicate a

natural rain in both droplet size distribution and intensity.

Rain intensities are relatively simple to control if droplet

size is disregarded. However, the HU0 scavenging rate per rain­

drop is not a function of rainfall intensity, but of the compli­

cated mechanisms of HCI-H20 mass-exchange-rates at the surface

which is influenced by raindrop size.

*Standard dry conditions

JIT RESEARCH INSTITUTE

10

ORIG ALpAG

o Poor QUALnm

IN

Figure 5

Rocket Motor Mountingwithin 5.49 m (18 ft) Spherical

Chamber.

Photo taken after firing. Note white Al 0 dust onupward facin aurfaces.

Rocket housing diameter is 7.5 cm (3 inches).

I 11

Engineer JET PROPULSION LABORATORY Batch No. LS-69 Charge No. 10

L. Strand LOADING DATA FOR CASE BONDED Job No. 365-40501-0-3450 INTERNAL BURNING TUBULARS Date Cast 7-22-76

OR 2-5071A STAR By Gurak-Ford-Tervet

PROPELLANT PLUGS

No. LS-69 Mixer No. ECCOBOND45

Viscosity__ Temp., OF Viscosity _ Temp., OF 16

Remarks: Remarks:

WEIGHTS (gms.)

1. Weight after adding end plugs (total weight). 1902

2. Weight after machining to length. 1896

3. Weight of plugs (1-2). 6

4. Weight of empty chamber.

5. Weight of propellant (2-4).

DIMENSIONS (inches)

1. Length of chamber. 6

*2. I.D. of chamber. 3

*3. Diameter of axial perforation (tubulars only). 2.042

*4. Diameter of inner star point circle (star only).

*5. Length of propellant after machining. 2.03

6. Length of propellant plus plugs.

7. Thickness of plugs (6-5).

TEMPERATURE HISTORY

MONTH 7

DAY 22

HOUR 12:30

TEMP., OF 1600

CURE 4DA

COOL 7-26

Figure 6

DATA SHEET FOR EXPERIMENTAL TEST ROCKET

12

Test Temp., OF 800 Run No.

Batch No: LS-69

Date 6/1/77 Charge No. 10 Job No. C6365

Observer D. Fenton JET PROPULSION LABORATORY Engineer P. Ase

SOLID ROCKET TEST DATA

Propellant No. LS-69-10

Purpose of Test Obtain rocket exhaust products in 18 ft. diameter space for study

Internal Tubular Star Special

Igniter Type JPL 540 Igniter Lot No. Wt., gms. 6.5 g

Squibs Nichrome wire

Safety Diaphragm: Metal Thickness, in. Dia., In.

Nozzle Type JPL-D9041893 4:1 expansion nozzle

Transducers PCB Model 113A24 nu5mv/psi Scope Tektronix 551

Power Supply PCB Model 482A, output neg. Calibration overall, 0.5 sec/cm

Charge Ampl. Kistler 556 - lmv/pcb Photo taken 2v/cm trace on scale-200 psi/cm

1v/cm trace off scale-lOOpsi/im

Remarks: Two traces were used with different sensitivity to catch one on scale.

Fir 6cn tinued)

13

The experiments evaluated the scavenging of HCl under mult­

iple raindrop sizes. A rain generation system was selected on

the basis of the capability to produce uniformly sized drops. The

method of raindrop generation was important as drop size enters

directly into the calculation for scavenging rate.

The raindrop generation system used is based on the capillary

instability principle for producing uniformly sized drops. This

system provides the advantage of control of both raindrop diameter

and initial fall velocity. The liquid jet break-up resultant of

capillary instability was theoretically developed by Lord Rayleigh

in 1878 (6). His development shows that a laminar, low-viscosity,

liquid jet is inherently instable due to surface tension and tends

to amplify longitudinal disturbances of a definite wavelength,

X, equal to 4.508 times the jet diameter, D.. As these waves

increase in amplitude, the jet breaks into equal-sized segments.

The resulting drop diameter, Dd, is therefore 1.89 times the

diameter of the jet. The equations for Rayleigh break-up are as

fbllows:

= 4.508 D.

Dd = 1.89 D.

F = VI/X

where F is the applied frequency of excitation and V. is the vel­

ocity of the emerging liquid jet.

Investigations relating to Rayleigh's work have determined

that the production of mono-dispersed droplets was achieved for

/Dj ratios varying from 3.5 to 7.0 (7, 8, 9, 10). This range

in the /D. ratio was determined from empirical measurements. J

The raindrop generator consisted of a square-ended (care­

fully machined) hypodermic needle mounted on an acoustic trans­

ducer. An ordinary audio oscillator was used to drive the trans­

ducer and therefore perturb the liquid jet. Figure 7 is a schem­

atic diagram showing the components of the raindrop generation

system. The water reservoir was a large polyurethane bottle

filled approximately to one-half the total volume with distilled

1IT RESEARCH INSTITUTE

14

HIGH AND MEDIUM FLOW SYSTEM

Compressed Absolute Air Regulator Filter Manometer

Signal Rain Generator Generator Water Reservoir

Figure 7

RAINDROP GENERATOR

lIT RESEARCH INSTITUTE

15

water. The flow rate of the rain water was regulated with the

use of carefully controlled compressed air. The complete genera­

tion system was located outside the experimental chamber to facil­

itate changing droplet size during scavenging tests.

Three raindrop sizes were generated during the performance

of the scavenging experiments. These sizes were 0.55mm, 1.1mm,

and 3.0mm in diameter and covered the operating range of the gen­

eration system used. Operational data for each of the raindrop

sizes tested are given in Table 1.

Preliminary experiments were conducted with the raindrop

generation system to evaluate overall performance. The genera­

tor was set-up in the laboratory under the desired experimental

conditions and the generated droplets were photographed using a

special optical system. A linear scale was located adjacent to

the falling droplets to directly determine size.

The 3.0mm drops were found to be too close together

(t 0.5mm gap) for adequate mixing with the rocket exhaust cloud.

As a consequence, a 18mm separation was achieved with a reduced -

Table 1 - liquid flow rate. Rayleigh break-up of the liquid jet

was clearly observed at the new flow condition and, therefore, en­

sured uniformly sized droplets. However, the droplet terminal

velocity was not achieved at the exit of the hypodermic tube. A

corrected time interval or residence time for the 3.0mm droplet

within the experimental chamber was measured and used to correct

the HCI scavenging data. The actual residence time was based on

16mm movie camera photographs of the falling drops inside the

experimental chamber. The camera was operated at a speed of 32

frames-per-second using Tri-X (Kodak) film. Through the use of

special lighting techniques, the raindrops were observed to pro­

duce visible tracks on the film. The number of frajes counted

for the complete fall of the drops varied from 22 to 26. This

gave an average residence time of 0.75 sec. for the 3.0mm drop­

lets. The calculation of the Hal scavenging rates incorporated

this pre-determined residence time.

The 1.1mm droplets were spaced approximately 2.5mm apart

and as determined from the photographs, were uniformly sized and

1IT RESEARCH INSTITUTE

16

Table 1

RAINDROP GENERATION SYSTEM CHARACTERISTICS

Raindrop Liquid Jet Needle Settling* Liquid Excitation Size Diameter Gage Velocity Flow Rate Frequency (mm) (mm) (Regular Wall) (cm/sec) (cm 3 /min) (HZ)

0.55 0.292 24 225 9.0 1705 1.1 0.584 20 434 69.8 1648 3.0 1.60 14 806(72.9) 88 101

*Based on work by Gunn and Kinzer (11). Velocity in parentheses is the actual droplet velocity upon droplet formation.

IIT RESEARCH INSTITUTE

17

falling at their terminal velocity. The droplet size measured

with the use of the scale was 1.1mm.

The generation of the smallest droplets was initially expec­

ted to occur in the 0.3mm range, but practical limitations neces­

sitated an increage in droplet size to 0.55mm. Due to the domi­

nating influence of the liquid's surface tension and imperfections

at the nozzle tip severely deflecting the liquid jet, 0.3mm drops

could not be reliably generated. Preliminary experimentation

determined that 0.55mm droplets avoided these problems and could

be generated at their terminal fall velocity. Separation distance

between adjacent falling droplets was approximately 1mm.

A commercially available fog nozzle* was secured and set-up

near the experimental chamber. The fog nozzle had a liquid rate

of 0.076z/min and required both a fixed gas pressure of 60 psig

and a variable liquid supply pressure of 5-30 psig. Stainless

steel was used in the construction of the nozzle. The advertised

droplet diameter was from 10 to 5 pm. The actual size of the drop­

lets was measured to be an average of 10 pm (based on number).

Visual appearance of the fog definitely suggested strong similar­

ity with naturally occurring fog. Time required to fill the

experimental chamber with a "thick" fog was on the order of 15

seconds -- waiting time was minimized. The fog nozzle itself

was kept outside the experimental chamber at all times except

when fog was introduced into the cahmber. At this point, the

nozzle was inserted through the open door of the experimental

chamber, permitting the fog to flow into the Teflon bag. A Milli­

pore filter was used in the water line of the fog nozzle to pre­

vent clogging.

2.4 Rain Collection

The rain collection system was located at the bottom of the

experimental chamber. The design of the collection system was to

prevent the "collected" raindrops from contacting the HCI exhaust

cloud above. This was crucial because the separate application

*Heat Systems-Ultrasonics, Model 700; New York, New York. lIT RESEARCH INSTITUTE

18

of a blank was avoided and the collected rain sample could be

analyzed directly.

Final design of the rain collection system is given in the

schematic diagram of Figure 8. A 20 cm (8 in.) diameter pyrex fun­

nel was used to direct the collected raindrops to a holding flask.

All fittings were ground-glass to minimize HCl losses. Nitrogen

was introduced into the flask and flowed upward forming a natural

barrier to the HCl cloud above: Exit ports were located along the

periphery of the funnel to remove the nitrogen. The flow through

the periphery ports was adjusted to match the flow rate of the

incoming nitrogen (14.2 1pm). The sheath formed by the nitrogen

trapped the exhaust cloud within the experimental chamber while

maintaining the cloud's natural concentration and decay rate. The

funnel assembly was sealed at the bottom of the chamber by a soft

rubber gasket which facilitated removal from its holder for clean­

ing between scavenging experiments.

A balnk of the rain water was obtained during each series

of scavenging experiments. The blank served as a correction to

the initial Cl- content of the rain water and was generally found

to be relatively low. Analysis for the Cl- content of the rain

droplets employed a chloride-specific ion electrode for the first

3 rockets ignited and colotimetric determination for the last two

rocket tests.

2.5 Measurement of Hydrogen Chloride Concentration

The measurement of the rocket exhaust cloud HCI concentration

-is important and has direct input to the raindrop scavenging data.

A meeting was held at NASA-Langley to determine the most suitable

detection method for HCI within the experimental chamber. The

people contacted at NASA-Langley were Dr. Gerry Gregory and

Dr. Scott Wagner of the Atmospheric Environments Branch. Agreed

at the meeting was the overall suitability of the chemiluminescent

HCI detector (Geomet)* for use in the IITRI experimental program.

*Geomet, Incorporated, 2814-A Metropolitan Place, Pomona, Calif.

SIT RESEARCH INSTITUTE

19

Nitrogen ExitRubber Gasket

Wood Holder

Nitrogen

250 ml Flask

Figure 8

RAINDROP RECEPTACLE

lIT RESEARCH INSTITUTE

20

NASA-Langley, at the early phase of the IITRI program, had experienced some difficulty with the gaseous phase HCI calibra­

tion procedure. The difficulty was in the passivation of the system to the low HC concentration levels--about 1 ppm. For

this reason, a liquid calibration procedure was recommended to

IITRI.

Two Geomet HC1 detectors were acquired for use during the program. The first instrument was purchased outright by IITRI

and the second was borrowed over the program's length from NASA-Langley. First, the instruments were calibrated--both at the initiation and conclusion of the scavenging tests to determine

and quantify sensitivity drift. Second, a technique was devel­oped and experimentally evaluated permitting the physical sep­aration of the gaseous and liquid phase HCl. The performance of the Geomet detector has already been documented by NASA-Langley, but careful calibration was nevertheless required.

The NASA-Langley "Field Calibration Procedures for Chem­iluminescent Hydrogen Chloride Detector" was used in the cali­bration effort at TITRI. The resultant instrument accuracy as a result of this method of calibration was approximately + 15

percent. The procedure called for the injection of a solution

of known HC1 concentration. The detector's response was then measured and related to the amount of HCl injected. Because the

HCI was in a liquid solution, a syringe was used to inject the liq­uid into the instrument's sample inlet tube. The composition of

the HC1 calibration solution is given below:

1. 4.9 ml methanol 2. 0.1 ml H20 3. 25 P1 constant boiling HCI azeotrope

The above fluid contained about 5.5 x 10-6 gm of HCI per 5 il

of solution.

The scale factor calibration--electronic gains between out­put scales on the instrument--were checked and found to be very near the nominal values (see Table 2) throughout the duration of

the program. The sample flow rates were also checked and set at 2.0 1pm. Modification of the IITRI detector was necessary to

SIT RESEARCH INSTITUTE

21

Table Z

SCALE FACTOR CALIBRATION FOR THE HCI DETECTORS

IITRI Detector Langley Detector

SFI* = 9.94 SFI = 10.14

SF=2 = 9.93 SF2 10.00

SF3 ***= 9.99 10.O4SF3

* SF1 -- scale factor for XI and X10

scales.

** SF -- scale factor for XIO and X100 scales.

*** SF -- scale factor for X0 and XIK scales.

lIT RESEARCH INSTITUTE

22

achieve the desired 2.0 1pm. Also, the inlet tube was recoated frequently to insure no depletion during calibration.

Liquid injections were made at the 3 V1 level to constrain the instrument response within the range of the recorder. Figures 9 and 10 show the typical traces obtained. The area under each response was calculated through the use of a Hewlett-Packard

9100B desk calculator and plotting board. At the beginning of the program, the IITRI detector exhibited a sensitivity of 10.85 ppm/volt on the XI scale while the Langley instrument gave 12.51

ppm/volt. This was considered to be typical of instrument-to­instrument variation. At the conclusion of the test program, the borrowed NASA-Langley detector had a sensitivity of 17.1 ppm/volt on the Xl scale while the IITRI purchased detector had a 19.3 ppm/volt sensitivity on the Xl scale. Application of the scale factor calibration yielded the sensitivity of the remaining scales. A significant shift in detector sensitivity occurred over the test sequence and, therefore, the data were corrected to compensate

for this shift.

A separate auxiliary experiment was conducted concerning the detection of airborne HCl. During the first set of IITRI tests (12), bubblers, employing distilled water as the collection liquid, were used to measure the HCI concentration during the scavenging tests. Questions arose revolving around the reliability of the bubbler technique. To check on this question, two midget impingers were connected in series, filled with 50 ml of distilled water each, and used to sample the experimental chamber's contents. When the chamber concentration was approximately 1.8 ppm HCI (1.82 ppm time-weighed-average), the bubbler train, operated at 4.9 1pm, indicated 2.0 ppm HCI (13 min). The chemiluminescent HCI detec­tor was taken to give the true HCl concentration and the associ­ated percentage error was calculated as 10%. This was considered reasonable in view of the low HCl concentration and gave credence to the earlier IITRI scavenging results conducted at significantly

higher HCI concentrations.

lIT RESEARCH INSTITUTE

23

Chart Speed = 1 mm/sec ( ) 2 cm/volt H0 IXTRI Detector

Figure 9

LIQUID CALIBRATION RESPONSE FOR THE IITRI HCI DETECTOR

Chart Speed = I mm/sec ( ) 2 cm/volt Langley HCI Detector

Figure 10

LIQUID CALIBRATION RESPONSE FOR THE LANGLEY HC DETECTOR

Depending on the "local" relative humidity, the combustion

generated HC1 might be in the gaseous phase or in both the liquid

and gaseous phases. This aspect had significant consequences

relating to the washout (below raincloud) of HC1. If a portion

of the HC1 was contained within the liquid phase, the washout

rate was reduced because of the lowered collision efficiency of

raindrops with the smaller droplets. However, during the initial

washout experiments conducted at IITRI, the expected large reduc­

tion did not occur under the predicted circumstances. Therefore,

improved experimental techniques were developed to provide details

of the HCl gaseous/liquid partition.

Several experimental schemes were explored, but only one

held promise. Before describing the technique utilized, the

others should be mentioned. Precipitating the liquid HC droplets

by means of an applied electrical field was considered because

collection efficiencies are good for submicron droplets (impac­

tion is not adequate for submicron droplets). However, an elec­

trical precipitator was attempted by Dawbarn (13) and failed to

function properly. An additional problem was the likely liber­

ation of the once liquid-borne HCl, thus providing an error in the

measurement of gaseous HC1. It appeared that this re-liberation

phenomenon could not be overcome. The same was true for direct

filtration; once the droplets were collected, they would evapor­

ate due to the passage of unsaturated air. No fool-pro6f scheme

could be devised to fix the HC1 (liquid phase) to the filter sub­

strate.

Whereas both of the above approaches removed the liquid

phase from the gaseous phase, it was also possible to trap the

gaseous phase HCl while allowing passage of the HC1 droplets.

Capillary tubes in a parallel configuration were found to provide

adequate opportunity for gaseous diffusion of HCI to the walls,

thus passing the liquid phase droplets. The material used to

construct the capillary tubes must fix the HC once contact was

made. Due to HC0's high chemical activity, a suitable surface

IIT RESEARCH INSTITUTE

26

was not difficult to locate--in fact, copper, brass, or stain­

less steel tubing were all candidates.

With the capillary tubes, the gaseous HC was lost and the

liquid phase HCl was introduced into the detector. A second HC

detector measuring total HC1 then provided sufficient detail to

determine the partition of the gaseous and liquid phase HC.

Subtracting the total HC concentration from the liquid phase

HC concentration gave the gaseous phase HC concentration.

Calculations have been performed to determine the per­

formance obtainable in fixing the gaseous HC by means of

capillary tubes. Sherwood, et. al. (14) gives the following

relation yielding the concentration of a gaseous species as a

function of axial length along the tube.

- 1 4 -8 9 " Ci - 0 , 22 +O.819e 63 0 + 0.0976e

C. ­0

where

= x

4r U00 C. = initial gaseous concentration

Co = final gaseous concentration

= average gaseous concentration

= diffusion coefficient

x = axial length along tube

r = tube radius0

U = mean flow velocity inside tube.

The results of an example calculation are given in Figure 11

showing the decrease in gaseous phase HC concentration with

downstream distance. The left-hand side of the above equation

can be referred to as the relative penetration. The calculation

assumes specific operating conditions that have been manipulated

to give good performance yet maintain simplicity of construction.

IIT RESEARCH INSTITUTE

27

100

droplet diffusion80

XNOTE: 5 tubes in parallel

tube dia. = 1 mm -o

0 ,m

gaseous HCI diffusion 4o

U

20

024 6 8 10

Axial Tube Length, cm

Figure 11

CALCULATION RESULTS FOR DROPLET AND GASEOUS DIFFUSION OF

HC1 SHOWING THEORETICAL MAGNITUDE OF SEPARATION

liT RESEARCH INSTITUTE

28

The tube diameter was 0.10 cm and the bundle consisted of 4 copper tubes. Because the sample flow rate into the HCl detector

was 2 1pm, the velocity.within the tube was not arbitrary, but 843 cm/sec. Calculations for the losses of droplets to the wall

require knowledge of size. Fuchs (15) gives for steady laminar

flow

= 0.891e-3' 65 7a + 0.097e -22 "3 -57a + 0.032e a + ni

where

a =4'x Uri

0

and4' is the diffusion coefficient for the droplets, n, the

mean number concentration of the droplets, and ni the initial

number concentration of the droplets. The influence of drop­

let size is made throughS-', which varies inversely with size.

If a droplet diameter of 2 pm is selected, the curve on Figure 11

is obtianed giving the relative number of droplets (penetration)

surviving the capillary tube. Larger droplets give greater sur­vival while smaller droplets give less. Gillespie and Johnstone

(16) obtained droplets grown in the presence of HCI gas to mean

sizes of 5.5 pm while the IITRI work on the HCl-H2 0 vapor systems

yielded droplets of approximately 0.7 pm growing rapidly to

larger sizes in a dynamic system (17). Since the exhaust mater­

ial within the test chamber was contained for a reasonable time

interval, the droplets grew providing the local conditions-were

correct (sufficient relative humidity).

To evaluate the workability of the diffusion battery con­

cept, a preliminary laboratory experiment was conducted. A

special air dilution system was constructed as shown in Figure 12.

The purpose of the dilution system was to deliver gaseous HCI

at a pre-determined concentration based on precise air dilution

flow rates. The method depended on knowing the initial concen­

tration of the HCI prior to dilution.

IIT RESEARCH INSTITUTE

29

TeflIon

DiffusionBattery

(4Copper Tubes)I.D.=I mm

Gls

_

0

W tf

no

-

z

Exhaust Hood

Flow Meter

m

Detector Detector

Recorder

2 0zeotrope

Figure 12

SCHEMATIC DIAGRAM OF AIR-DILUTION SYSTEM

Clean Dry

Air

A commercial high pressure tank of HCl gas diluted with

nitrogen to 998 ppm HC1 (certified) was purchased to provide

the known source of HC1. However, when the total dilution sys­

tem was operational, the HCI content was drastically reduced as

indicated by the Geomet detectors. Humid air was introduced to

the dilution air with no change in HC1 response. Therefore, it

was concluded that the HCl in the tank had probably reacted with

the metallic walls. As indicated in Figure 12, the HC1 bottle

(tank) was replaced with the HC-1H20 azetrope.

A larger flow rate (8 1pm) was passed through the diffusion

battery (15.2 cm long). Carrying out the calculations based on

gaseous diffusion mechanisms, the HCI concentration ratio was

0.39 while the measured ratio was 0.5. This ratio remained con­

stant for approximately an hour indicating the expected "life"

of the copper tubes. Concluded, therefore, was the workability

of the copper diffusion battery concept for the separation of

the gaseous and liquid phases of HC1.

2.6 Absorption Chamber

The absorption of 1CI by terrestrial surfaces such as sea

water and the ground could significantly affect the HCl concen­tration predicted by the MSFC multi-layer diffusion model. The

MSFC model, at the present stage, assumes no absorption at the

surface but can be easily modified to account for HC1 loss by a

normalized "absorption" coefficient.

During the absorption chamber experiments, four surfaces

were evaluated for relative HC uptake. The surfaces tested

included distilled water, sea water, brackish water, and a refer­

ence solution of NaOH (0.05 Normal). The reference solution

serves as the normalization surface for the HC1 uptake. The

composition of sea water was privately obtained from Mr. Chao

Schem at the Naval Air Propulsion Test Center (18) and is given

in Table 3.

The composition of brackish water cannot be precisely spec­

ified. In fact, the term "brackish" is very general and incor­

lIT RESEARCH INSTITUTE

31

Table 3

COMPOSITION OF LABORATORY SEA WATER (18)

Sea Water: 1 liter

NaCI 23 g

NaSO4 * 10H20 8 g

stock solution 20 ml

distilled water

Stock Solution: I liter

to yield I liter

KCl

KBr

0g

45 g

MgCl2 26H20 550 g

CaCl2 • 6H20

distilled water

110 g

to yield 1 liter

32

porates the marshy backwater of rivers, the bay areas near the

sea, and industrial effluents. Consequently, the brackish water

most desirable for these absorption experiments is that water

from the Cape Kennedy area. However, our experiments were not of

sufficient precision to warrant the expense of obtaining specific

brackish water. A "working" definition of brackish water was

obtained from the Environmental Protection Agency (Region 5)

and was applied here--40% solution of sea water with distilled

water.

The normal procedure for calculating the removal rates of

gaseous pollutants involves the so-called deposition velocity.

The deposition velocity is defined by F=Vc

g

where F is the downward flux of the gas and c the concentration

of the gas at a specified height above the surface where the

deposition occurs. Another approach utilizing the concentration

gradient immediately above the surface and the local atmospheric

diffusivity can also be used to determine deposition velocity

but was unsuitable for the present experimental facility. Mea­

surement of the turbulent diffusivity would be much too cumber­

some. Therefore, V was determined as given in the above sim­g plified equation.

The use of deposition velocities rather than total fluxes

was desirable because the data are automatically corrected for

differences in HCl concentration. The experimental data obtained

gave the HCl flux, F, and an average HCI concentration measured

downstream from the absorbing surface. Concentration values

were also checked upstream and were found only slightly higher

than the downstream values.

Figure 1 shows the general location of the absorption duct

within the overall experimental facility. The rocket exhaust

cloud must pass through the duct assembly upon transport to the

experimental chamber. The dilution system, with its ability to

IIT RESEARCH INSTITUTE

33

provide excess air at varying humidity, served to regulate the

volume flow of rocket exhaust from the spherical holding chamber.

The design of the HCl absorption duct is given in Figure 13.

Galvanized steel sheet was-used to form the walls of the duct.

The inside surface was coated with an epoxy-resin paint to mini­

mize HCl loss and subsequent corrosion of the inside surface. Two

and one-half cm (1 in.) pipe fittings comprised both ends of the duct

:and facilitated connection to the remainder of the experimental system.

The transition from the 15 x 10 cm (6" x 4") rectangular duct to

the one inch pipe fittings was 15 cm (6 in.) long and rectangular

in cross-section. The total length of the straight portion of the

duct was 60 cm (2 ft.). A tray was provided for the absorption

experiments which was 5 cm by 30 cm (2" x 12") and approximately

2.5 cm (I in.) deep. A gasket was used to provide a tight seal

between the 2.5 cm (1 in.) flange on the tray and the bottom of

the duct.

Basically, the experimental procedure was to measure the HC1

picked-up by various surfaces through analysis of the surfaces

themselves for chlorine. Knowing the HCl concentration, velocity,­

and exposure time would yield the "absorption rate", deposition

velocity, and the mass transfer coefficient. The term "absorption­

rate" is not precisely correct because absorption is a specific

local mechanism and separate from HC1 transport to the surface

where absorption may or may not occur.

A question arose regarding the fact that the experimental

procedure did not measure the actual depletion (or depletion

rate) of the passing HC exhaust flow. Therefore, the suggested

experimental plan was to measure both the inlet and outlet HC

concentration and by arithmetic difference, calculate the rate

of HCI depletion. An initial inlet and outlet measurement was

required with the absorption chamber empty in order to "calibrate"

the wall losses. Therefore, the difference between the calibra­

tion test and those with various absorbing materials gave the

depletion rate of HC1 for the specific materials exposed

IIT RESEARCH INSTITUTE

34

15 cm 60 cm 15 cm-­

-II

Mi M

LUU, z

z 1 inch 2.5 cm ~pipe fitting

30 cm

-2.5 cm flange tay for experiments

Figure 13

DESIGN OF DUCT FOR HCI ABSORPTION EXPERIMENTS

Note: 1. Duct cross-section is 15 cm x 10 cm. 2.* Tray inside dimensions is 30 cm x 5 cm x 2.5 cm deep.

To check the workability of the suggested procedure, a

preliminary experiment was conducted. Figure 14 shows a schematic

diagram of the test equipment. The absorption chamber used was

the same as constructed for the HCl abosrption experiments. A

supply of clean air (regulated and filtered) was passed through an HCl-H20 azeotrope to deliver gaseous HC1. Shop air, monitored

with an inclined manometer, provided the main air flow through

the absorption chamber. A dry-test meter was used to measure

total flow rate and calibrate the flow system.

Table 4 gives the results obtained for both the chamber

calibration runs and the distilled water test runs. All the

tests were conducted in the HCl concentration range of al ppm.

Noted immediately was the consistent reduction of the HCl

concentration upon passage through the chamber. However, an

explanation cannot be offered as to why the distilled water tests

yielded a lower HCl concentration difference than the empty cham­

ber tests. The difference should have been greater due to the

presence of the distilled water (HCl sink). Consequently, the

reliability of the H0 concentration measurement may be suspect.

The calibration error in the HC1 detectors is in the vicinity of

%15% on the 10 ppm HC1 scale (based on the calibration previously

performed). Therefore, the uncertainty in the HC1 measurement is

about + 1.5 ppm on this scale. The differences in HC1 concentra­

tions measured by the two HC detectors are well within this

uncertainty and are, therefore, not reliable. This does not mean

that the HC detectors are not suitable for general measurement

of HC1 concentration levels, but only that their application here

was unwarranted due to the precision required.

Further, the distilled water was retrieved and analyzed for

Cl content with the use of an ion-specific electrode. These

results are also shown in Table 4 where a steady increase in

C1 content of the distilled water is observed with increasing

mean velocity within the chamber. Measurement of the C- content

of the distilled water was strictly routine. Thus, the deter­

mination of the HC mass flux to the surface was readily available.

IIT RESEARCH INSTITUTE

36

Incline Manometer.

Absorption Chamber Outlet

Liquid Holding Tray

- N Dual Channel HC1 z Detector Recorder Detector

S(Geomet) (Geomet)

Clean Air Supply

HCl Azeotrope Supply Air Pressure

Figure 14

SCHEMATIC DIAGRAM OF EQUIPMENT USED IN PRELIMINARY HC1 ABSORPTION EXPERIMENT

Table 4

PRELIMINARY TEST RESULTS FOR ABSORPTION CHAMBER

Time Measured Total Interval Mean Velocity Cl- Content** Inlet HCI Conc.

Chamber Calibration min. (cm/sec) (mg) (ppm)

(stable) 3.97 0.98

(stable) 5.21 1.23

(stable) 7.30 1.48

Chamber Test with Distilled Water

1.0.8 3.97 0.32 0.98

9.2 5.21 0.54 1.35

9.9 7.30 1.1 1.35 Go

* Based on cross-sectional area of absorption chamber **Total volume of distilled water 200 ml, distilled water -- 0.22 mg Cl

Outlet HCI Conc. (ppm)

0.545

0.76

1.10

0.82

1.20

1.31

HCI Conc. difference

0.44

0.47

0.38

0.16

0.15

0.04

Figure 15 displays the three data points in terms of the mean

HC1 fluxes through the absorption chamber and into the distilled

water.

As a consequence of this preliminary absorption duct experi­

ment, the employed methodology of HU1 detection was not suitable

for absorption rate measurements based on differences. The required

accuracy and precision of measurement was beyond the instrument's

calibration errors. Therefore, the direct measure of C1 within

the absorbing media was performed for each of the uptake tests.

2.7 Growth Chamber for Experimental Plants

Selection of the plant material to undergo HC1 uptake exper­

iments was made at the beginning of the program. Before repre­

sentative samples of the flora in the area of the Shuttle launch

site are monitored, experimental plants serving as standard

laboratory subjects, where background Cl- measurements have

already been perfdrmed under numerous conditions, should be eval­

uated. This was the approach taken with the experimental program

conducted here.

The two standard representative plant material types are:

monocotyledonous plants (oats and corn)

dicotyledonous plants (peas and beans)

The actual plants selected were corn-and soybeans for the practi­

cal reasons of easy growth and care.

A "growth chamber" was built to enable controlled plant growth.

This chamber permitted control of the humidity, temperature and light

that the plants received, thus removing some of the variables nor­

mally present. The growth chamber (Figure 16) consisted of the

growing area approximately 2.43m x 1.83m x 1.23i (8 x 6' x 4'), a

humidifier, blower, heater, and light tray. The light tray had

twelve F96T12 high-output fluorescent tubes and six 150w/130v incan­

descent lamps to simulate the light spectrum produced by the sun.

For optimum growth the plants require about 2000 ft-candles of

light. To allow for growth, the height of the light tray was ad­

justed to maintain the 2000 ft-candles intensity. IIT RESEARCH INSTITUTE

39

0.12 7

0.10

0.08

o

)

0.0

0.04

fa 0.02

m C

0 0.04 0,08 0.12 o.16

Mean HCI Mass Flux Through Chamber (mq H)

Figure 15

-SURFACE ABSORPTION FLUX OF HCL TO DISTILLED WATER VERSUS MEAN GASEOUS HC1 FLUX ABOVE SURFACE

[IT RESEARCH INSTITUTE

40

Greenhouse

57"

>

-4

MI" n

z A

I 103 1-

Light Tray

4711

Fluorescent Lamp

Incandescent Lamp

4411

L7"

Figure 16 PLANT GROWING CHAMBER

The photograph shown in Figure 17 was taken about 3 weeks

after planting the seeds. Both corn and beans were raised ­

the beans are behind the corn and are only partially visible.

The growth achieved demonstrates the growing capacity of the

chamber.

The plant watering system was made automatic. Tygon tubing along the upper edge of the plant box had small holes that pro­

duced small jets of water when the line was pressurized. The jets

possessed sufficient strength to penetrate all the plants. A

time interval of 15 seconds was found to keep the soil in a

moist state. Timers were used to regulate the time of plant

watering (only morning), the duration of watering, and the time period of lighting (12 hrs.). This was considered necessary to

minimize the growing variations from one set of plants to another

Also, an adequate supply of seed was obtained to insure that all

the seeds originated from the same source. The large conduits in the foreground of Figure 16 were the inlet ducts to the growth

chamber from the humidifier providing cool mist air.

Preliminary experiments were carried-out to ascertain the

number of plants needed to give a measurable chloride-ion back­

ground count. When using corn only 5 plants were required, but

with soybeans 15 plants were necessary. The chloride-ion quan­

tity at those numbers was easily measured by the chloride-ion

electrode method. The time of growth for a useable sample for

exposure in the experimental chamber was approximately two weeks after planting. Homogenization of the plants before analysis

was done with the plants dried. In this way the plants could be stored without loss or change of the chloride-ion content until

the analysis for Cl was performed.

2.8 Metallic Coupons

The corrosion of structural surfaces at the Kennedy Space Center, as well as those associated with private individuals

(i.e. automobiles and homes), due to the presence of HCl within

the ground exhaust cloud, were evaluated through the use of metallic coupons. In addition to the normal corrosion eff cts,

liT RESEARCH INSTITUTE

42

ORIGINAL PAGE lb OF POOR QUALY

m

Ii

Figure 17

CORN AND SOYBEANS (BEHIND CORN) RESPONDING TO

CONDITIONS INSIDE GROWTdH CHAMBER

lIT RESEARCH INSTITUTE

43

premature failure of a structure might occur through stress­

corrosion cracking. With stress-corrosion, any stressed metal

or alloy could undergo failure at stresses below the normal frac­

ture stress.

Test samples or coupons were fabricated from test alloys.

Table 5 gives the alloy-types comprising the test matrix where

each coupon was 2 inches square and 1/8 inch thick except for

aluminum siding (building construction). The edges of each

coupon were ground to a well-rounded profile to prevent any unus­

ual or promoted corrosion in this region. In addition, stress

corrosion specimens were fabricated from mild steel into a "U"

shape with a bar welded across the open end of the "U". Mild

steel was used for the painted coupons because automobile bodies

and structural support systems are generally mild steel. One

coupon was painted with a red lead paint while the second was

painted with a typical automobile finish. A second austentic

stainless steel coupon was coated with a special paint for space

vehicles (S-13G). Therefore, a total of 9 test coupons comprised

the matrix exposed to the rocket exhaust.

After exposure to the HC rocket exhaust, the coupons were

placed in a controlled oven maintained near saturated conditions.

The temperature was constant at 95°F and a pan of water at the

bottom of the oven maintained the high humidity. In this way,

long-time exposures of the exposed matrices were possible. Weight

differentials occurring from the exposure indicated the magnitude

of the surface corrosion present.

3. EXPERIMENTAL PROCEDURE

The experimental procedure used during the tests is described

here. Where required, clarification is provided after the specific

procedure is identified.

1. Test chambers were prepared for rocket igiition. This included the dealing of the large spherical chamber to prevent loss of the rocket exhaust cloud. The experi­mental chamber was isolated (valve) from the contents of the large chamber. Experiments performed inside the large chamber were installed prior to sealing and rocket ignition.

IIT RESEARCH INSTITUTE

44

Table 5

MATERIALS FOR CORROSION TEST MATRIX

Aluminum: 2024

6061

Steel: mild steel

Austentic Stainless Steel: 304

Painted: mild steel; red lead paint mild steel; automobile finish space paint

Aluminum siding: coated

IIT RESEARCH INSTITUTE

45

2. The rocket was ignited.

3. After approximately one minute had elapsed, the test plants and metallic coupons were installed within the experimental chamber. The rocket exhaust was drawn into the chamber until the desired concentration was achieved. The mixing fan was operated continuously to ensure uniform composition within the chamber. The decay in HCI concentration was compensated by adding more rocket exhaust as the exposure continued.

The exposure time was kept constant throughout all the tests at

20 minutes. This length of time is typical of the expected expos­

ure time for the actual shuttle ground cloud. Lights were installed over the experimental chamber to maintain the test plants at the

-same conditions within the growth chamber.

4. Upon completion of the exposure period, the rocket exhaust was first purged from the experimental cham­ber. A record of the HCl concentration and relative humidity was obtained and the time-averaged values used to characterize the exposure (the fog nozzle was employed to increase the humidity when required). Retrieval of the plants and metallic coupons took place with the plants cut about I inch above the soil sur­face and-dried immediately.

5. The rain scavenging tests were then performed. Again, with the application of the fog nozzle and alternate charging and bleeding of the experimental chamber, the desired test conditions were achieved. The raindrop generator was then started. During all the scavenging tests, the nitrogen flow to the rain collection system was operating. After a reasonable volume of raindrops was collected, the particular scavenging test was com­plete. The collected rain was removed and put in stor­age, the collection apparatus cleaned, and the rain generation system changed (exitation frequency, hypo­dermic needle, and liquid flow rate) before the next scavenging test was begun. All the scavenging tests were performed within this period of the test sequence.

6. The last series of tests involved the absorption duct assembly. The exit of the duct was monitored for HCl concentration. Valves were arranged in such a way to connect the absorption duct directly to the laboratory's negative pressure. In conducting a test, the lower tray was removed, cleaned, filled with the test liquid, and then re-installed.- The rocket exhaust was then caused to flow through the chamber and diluted to give the desired HCI concentration. The exposure time period was on the order of 10 minutes. After the test was

lIT RESEARCH INSTITUTE

46

complete, the tray was removed and the exposed liquidplaced in storage for later analysis. Preparations were then made for the next test liquid.

7. At the conclusion of the test sequence, the large cham­ber was ventilated and the experimental chamber manually washed clean. Experiments conducted within the largechamber were now retrieved. Once ventilation was com­plete, the large chamber was washed-down with the inter­nally located spray nozzle

Once the chlorine analysis was completed and the test.results reviewed, preparations were made for the next rocket firing.

4. EXPERIMENTAL TEST DATA AND RESULTS

4.1 Scavenging Test Results

The raindrop scavenging experiments were conducted over a wide range of HCI concentrations (0.2 to 10 ppm). Three rain­drop sizes were selected to embrace the size range for naturally occurring rain. These raindrop sizes are 0.55mm, 1.1mm, and

3.0mm. Within this section of the report, the scavenging data is both presented and reduced to an empirical relationship.

4.1.1 Rocket Exhaust HCl Scavenging Data

The HC scavenging data obtained with the experimental apparatus already described are given in Table 6. The relative

humidity within the experimental chamber was determined for each scavenging test (battery operated dry-bulb and wet-bulb thermom­

eter system). The variable, RA, is the rate of HC1 absorption by a single droplet of specific diameter and has the units--g/sec. The reported datum for the 0.3mm droplet is not considered reliable

because difficulty was experienced with the rain-generating equip­ment. The remaining data, with the same experimental techniques and procedures applied to each, are considered reliable.

In reducing the experimental scavenging data, a parameter

was sought to generalize all the data. A dimensionless-group

of variables can be formed as RA/PVsDd2 , where Vs is the terminal fall velocity for a particular droplet of diameter Dd. To force

the units to .cancel, the density of air, p, is introduced to the parameter. Plotting the parameter against the corrected HCI con­centration (corrected for HCI detector drift), Figure I8 results.

ItT RESEARCH INSTITUTE

47

Table 6

DROPLET SCAVENGING RESULTS FOR ROCKET EXHAUST HCI

Uncorrected Reference Corrected

Rocket Size Vol. Total Cl

Raindrop C1 Cone.

Raindrop C1 Cone.

Raindrop Cl Cone.

Corrected HCl Cone.

Experimental Chamber R.H. RA

Number (mm) (ml) (g) (pg/mi) (pg/Mi) ( g/mi) M M) (g/sec)

1 0.3* 14 "600 42.8 10.9 31.9 7.0 u55 3.8 x 10 -

1.1 152 4080 26.9 10.9 16.0 7.4 I55 3.5 x i0-7 3.0 152 3000 19.7 10.9 8.8 8.0 IU55 1.8 x 10

2 0.55 50 220 4"4 3.0 1.4 1.7 76 2.4 x 10­7

0.55 44 1000 22.7 3.0 19.7 8.4 62 3.6 x I0-F61 1.1 1.1i 3.0 3.0

155 150 150 160

-520 1240 520 600

3.4 8.3 3.5 3.8

3.0 3.0 3.0 3.0

0.4 . 5.3 0.5 0.8

1.9 9.4 2.0 9.4

95 65 '90 95

8.7 x 10-10 1.2 x 108 1.0 x I0­1.7 x 10 -8

3 0.55 54 320 5.9 3.3 2.6 1.5 90 3.7 x 10-10 0.55 66 820 12.4 3.3 9.1 9.0 43 1.3 x 00-9^ 1.1 170 600 3.5 3.3 0.2 1.8 84 4.2 x 10- 8 1.1 180 1680 9.3 3.3 6.0 9.0 78 1.3 x 10­3.0 3.0

150 180

800 1320

5.3 7.3

3.3 3.3

2,0 4.0

2.0 9.5

90 76

3.9 x 10 7.7 x 10- U

4 0.55 70 39.9 '0.57 0.03 0.54 0.19 73 3.1 x 10_1 0 0.55 50 347 6.94 0.03 6.91 2.5 35 9.8 x 1.1 1.1 3.0 3.0

145 155 165 145

4.35 205

4.95 55.1

0.03 1.32 0.03 0.38

0.03 0.03 0.03 0.03

0.0 1.29 0.0 0.36

0.25 1.8 0.18 1.9

35 35 32 32

---­2.7 x 10-9

---­2.7 x 10 - 9

5 0.55 35 21 0.6 0.0 0.6 0.47 89 8.6 x 10-11

0.55 48 24 0.5 0.0 0.5 0.69 58 7.1 x 10­10 0.55 76 175 2.3 0.0 2.3 1.7 61 3.3 x 10-10 1.1 134' 0.0 0.0 0.0 0.0 0.44 96 1.1 175 0.0 0.0 0.0 0.0 0.63 66 .. 1.1 156 125 0.8 0.0 0.8 2.0 60 1.7 x 10 9

1.1 .1

170 160

238 160

1.4 1.0

0.0 0.0

1.4 1.0

2.0 2.1

58 66

2.9 x i0o 2.1 x 109

1.1 142 170 1.2 0.0 1.2 2.3 66 2.5 x 109 1.1 145 203 1.4 0.0 1.4 3.1 58 2.9 x 10 8 3.0 128 88.9 0.7 0.0 0.7 0.86 91 1.3 x 109 3.0 3.0

140 122 .

28.0 12.2

0.2 0.1

0.0 0.0

0.2 0.1

0.86 3.3

91 62

3.9 x 10 1.9 x 10 -9

*Unreliable

- 4 , t1ilO

A

- 6

ixlO

RA--­ = (2.82 x 10)

032-5 (6.2c72 x 1 82'

PV3D

i Droplet Size

A Dd= 0.55 mm

IXI0 -7 A Dd= 1.1 mm 0 Dd= 3.0 mm L,Dd 0.9 nmm,

Knutson and

0 Open Symbols: RH < 80% Fenton (2) O Closed Symbols: RH>80%

1xIO0 8 * 0 1 0.2 0.5 1.0 2.0 5.0

I 10.0

i ?0.0 50.0

I 100

I 200 500 1000

Corrected 1CI Concentration (ppm)

Figure 18

ROCKET EXHAUST HCI SCAVENGING DATA FOR MULTIPLE RAINDROP SIZES

The parameter is observed to generalize the scavenging data

fairly well.

Error limits cannot be calculated for the scavenging data

as simultaneous "blank" experiments were not conducted. However,

one particular set of test conditions was repeated four times to

determine precision within the experimental technique (rocket

test number 5--l.lmm droplet diameter). *The percent deviations

are noted to be approximately 35% and may vary for the other two

droplet sizes as the typical chlorine concentration levels vary.

For the larger 3.0mm droplets, lower chlorine levels should mean

reduced experimental precision while for the smaller 0.55mm drop­

lets, experimental precision is probably enhanced.

Based on theoretical analysis, the relative humidity was

thought to strongly influence the scavenging rate. The theory

asserted that droplet growth was accelerated by the presence of

HCI and, therefore, once cloud nucleation occurred, the HCI con­

tent should have been predominantly in the liquid phase for rel­

ative humidities greater than 80%. At this point, the scavenging

rate was reduced because of the lower collision efficiency. In

contrast to this assertion, the experimental data displayed no

strong influence of R.H. on the HCU scavenging rate. High humid­

ity did show a slight depression in scavenging rate, but the

depression was within experimental errors. Further work relating

to the influence of the rocket exhaust cloud constituents upon

scavenging pure gaseous HCI is presented in the appendix.

Several scavenging experiments were conducted in the pre­

sence of fog. The experiments were conducted as follows. The

experimental chamber was first filled with a visible fog gener­

ated with the commercial fog nozzle. Initially, the HCl detec­

tors indicated that the HCl was predominantly gaseous. After

approximately 10 seconds, the gaseous HCI concentration decreased

until the dominant HCI phase was liquid HCI. This experiment was

repeated several times during three rocket tests with the same

results--virtually total transformation of the HCI to the liquid

phase over a very short (5-15 seconds) time interval.

IIT RESEARCH INSTITUTE

50

Droplets formed on the inner wall of the experimental cham­

ber were checked for pH after approximately two minutes of expos­

ure to the exhaust cloud (A2.O ppm HCl). The pH determined by

sensitized paper was about 3.1. The pH value was significantly

greater than the unity value reported during field measurements

by NASA-Langley. A plausible explanation is that the field mea­

surements collected droplets grown in size promoted by the pres­

ence of HCl-mist formation. This explanation is also likely

from consideration of the close proximity of the rain collection

site to that of the launch vehicle.

4.1.2 Correlation of Scavenging Data

Due to information concerning the terminal settling velo­

city of raindrops as a function of diameter, an empirical cor­

relation can be deduced from the experimental scavenging data.

This section discusses the calculations leading to the final

correlation.

Marshall and Palmer (19) presented an expression for rain­

drop concentration and size as follows:

Crd 0.Oexp 21j

where

Crd = concentration of raindrops of diameter Dd to Dd + ADd drops\

Dd = raindrop diameter, (cm)

R = rainfall intensity,(m()

Dingle and Hardy (20) have evaluated the Marshall-Palmer expres­

sion for a number of raindrop size spectra and have determined

the general validity of the expression on the average. Some

individual rainfalls were noted to vary widely from the norm.

Gunn and Kinzer's (11) data were used for the terminal

settling velocity of the raindrops. Figure 19 shows the dis­

crete points comprising the data and a power curve selected to IIT RESEARCH INSTITUTE

51

1000

900

0

0 n

7000

o 600

~s 0 V 1671D 0. 6 8 2 161d

P4

'-I0

0

500-C

/S

200­

100 ­

30

0 I! I

0 0.1 0.2 0.3 0.4 0.5 0.6

Raindrop Diameter, Dd, cm

Figure 19

RAINDROP TERMINAL FALL VELOCITY AS DETERMINED BY GUNN AND KINZER (11)

52

0.7

fit the data where the coefficient of determination was 0.93--a

relatively good fit. The terminal velocity data was not con­

nected for temperature, relative humidity, or local wind velocity.

The HCl scavenging data presented in Figure 18 is fitted by

a power curve as follows:

RA -= (6.72 x 10­ 5 ) MA0.824

pVD A

where MA is the mass concentration of HCl (g/m3). Ddfining Q

as the rate of removal of HCl from a cubic meter of exhaust cloud,

it then follows that

SRA(Dd Dd

) Crd (Dd ) &Dd.

Now, defining the washout doefficient, A, as the HCl removal rate

over the HCl concentration, A then becomes

A= MA RA(Dd)Crd (Dd) &Dd.

Substituting the experimental correlation and the power func­

tion for the raindrop terminal velocity, the follwoing.integral

=(.8resultsrKAeut j-3) 0.176 2.682 [- 4l'Dd] -3) A = (8.98 x 10 M 6 D exp UDDd D 2lR d.

Carrying-out the integration utilizing numerical procedures

(increment size for Dd = 0.04 cm) for several rainfall intensities

varying from 1.0 to 30.0 mm/hr. yeilds

0 773 A = (4.21 x 10-

8) PMA0.176 R

and under standard atmospheric conditions for p gives "

A = (5.12 x 10 5 ) MA0.176 R 0 7 73

IIT RESEARCH INSTITUTE

53

The above empirical relationship is different from the

earlier reported relations (1, 2)

56 7A = (8.3 x 10-5 ) R0 . , Knutson and Fenton (2)

4) R0 "6 25 A = (1.11 x 10- , Pellet (1)

by virtue of the non-linearity. However, the approach taken here

to deduce the HCl washout correlation is based strictly on the

data obtained. Both previous investigations assumed the form

of the Frossling correlation and fitted the data accordingly,

hence the perfect linear correlation. The fact that the expon­

ent of MA is relatively small--0.176--the influence of MA on A

is weak over the HCl concentration range tested. The total influ­

ence of the HCI concentration on the washout coefficient is only

about a factor of 2.1 and is seen to be within the errors incor­

porated within the experimental data. In this sense, the relation

found here supports the earlier correlations.

4.2 Absorption Tests

The absorption of rocket exhaust HC1 was measured for sev­

eral liquid surfaces typical of the region near Cape Kennedy. The

liquids tested included sea water, brackish water, distilled water,

and a 0.05 normal solution of NaOH as a reference. The composi­

tion of these liquids and the procedures used during the exper­

iments was described earlier within this report.

Table 7 lists the collected data and the calculated results

for the deposition velocity. Immediately apparent from the data

is the extremely large deposition velocities and HCl mass fluxes

into the sea and brackish water surfaces. These values are unre­

liable because of the high chlorine-ion background values (inher­

ently present for these liquids) coupled with chlorine-ion analysis

techniques of insufficient sensitivity. As a consequence, experi­

ments with both sea and brackish water were dropped from the test

sequence.

Reviewing the data from only the distilled water and the

0.05 noraml solution of NaOH, the deposition velocities (and HCl

mass flux values) are reasonable, relative to comparable gaseous lit RESEARCH INSTITUTE

54

00

Table 7

ROCKET EXHAUST 1CI ABSORPTION TEST RESULTS

Rocket Number

Exposed Material

Vol. (ml)

Time Interval (ain)

HCI* Cone. (ppm) (mg/m)

Reference Cl Cone. (pg/ml)

Uncorrected C17 Cone. (Pg/ml)

Corrected Cl- Cone. (pg/mi)

HC1 Flux F

(g/m2 see)

Deposition Velocity, Vg

(cm/sec) Normalized

yg 2 Distilled H20 200 10 24 36 2.1 5.5 3.4 8.79 x 10-5 0.24 0.22

Brackish H20 236 10 30 45 13,900 13,200 -700 -- --

Sea H20 250 10 27 41 25,900 40,000 14,100 2.56 x 10-1 1100 1000

Reference NaOH 248 10 26 39 18.9 32.9 14.0 4.48 x 4 1.1 1.0 (0.05 Normal)

3 Distilled H20 230 10 26 39 2.4 7.2 4.8 1.14 x 10 - 4 0.33 0.31

Brackish H20 280 10 31 48 11,100 13,900 2,800 7.78 x 10- 2 180 170

Sea H 0 2

Reference NaOH

260

190

10

10

29

29

45

45

30,200

20.6

40,400

42.7

10,200

22.1

2.63 x 10 - I -4

4.17 x 10

660

1.1

610

1.0 (0.05 Normal)

4 Distilled H20 200 30 7.9 11.9 0.015 2.52 2.50 1.90 x 10-5 0.21 0.78

Reference HaOH 235 30 7.1 10.7 1.60 4.17 2.57 2.27 x 10-5 0.27 1.0 (0.05 Normal)

5 Distilled H20 232 20 10.2 15.8 0.0 4.1 4.1 4.08 x 10-5 0.35 0.84

Reference NaOH 275 20 8.8 13.0 0.5 4.0 3.5 4.14 x 10-5 OA1 1.00 (0.05 Normal)

*Time averaged values over time interval for absorption experiment.

pollutants. Moreover, the normalized (reference solution) depo­

sition velocity ratio is reduced for the high HCI concentrations. The reduction in the deposition velocity ratio corresponds in

magnitude to the change in normalized HCl concentration.

Also note, as shown in Figure 20, that the HCl deposition velocity for NaOH is concentration dependent, while the deposition velocity for H20 is not dependent on the concentration. An explan­ation for this may be that since the absorption surface is stag­nant, the liquid immediately below the free surface becomes satur­ated at the higher HCl concentrations more readily, thus restirct­

ing continued absorption of HCl. With the lower HCI concentration

above the free surface, longer exposure times are required for

saturation to occur therefore'absorbing a relatively larger amount of HCI compared to the reference surface over the same time interval.

4.3 Plant Uptake Tests and Growth Response

The exposure of standard test plants, a monocotyledonous ­corn, Zea mays, and a dicotyledonous - soybean, Glycine max, were

used to determine the amount of HCl uptake during a specified period of exposure. A preliminary test was conducted at the con­

clusion of the experimental test sequence to determine real-time

growth response.

Related work has been recently reported or is under way in relation to plant injury caused by exposure to the rocket exhaust.

The Kennedy Space Center has sponsored investigations on plant injury both at the Florida Technological Tnstitute and North

Carolina State University. The Florida work used scale-model

rockets to generate an exhaust cloud within a Teflon tent enclos­ing the vegetation of interest near the launch site. Brown spot­

ting on the plants was observed but the HCl-concentration measure­

ment was considered unreliable (21).

The work at North Carolina State University is being currently

managed by Dr. Walter Heck (22). Experimental results are not

available from this study. The experiments are planned to eval­uate injury to a number of plant species within specially designed

IIT RESEARCH INSTITUTE

56

1.2

1.1 230 E

1.0

u 0.9

r4 0.8 00 Distilled H20

.

0 W

0.7

0.6 El Reference NaOH (0.05 normal)

0•

S

0m

p

0 .5

0.4

0.3dEl

-

30

0.2 0

0.1

0 0 5 10 15 20 25

Corrected HC1 Concentration (ppm)

Figure 20

30 35 40

ROCKET EXHAUST HC1 DEPOSITION VELOCITY VARIATION WITH HC1 CONCENTRATION

exposure chambers. Lerman et. al. (23), while at the Univer­

sity of California at Riverside, exposed numerous plant species

to pure gaseous HC1 and found Geraniums to be most sensitive to

HCt exposures.

The experiments at IITRI, therefore, were not directed

toward the determination of plant injury, but rather, toward the

uptake of HCI. This concern stems from the requirement for re­

liable HC1 "sink" terms comprising the multi-layer diffusion

model describing the dynamics of the ground exhaust cloud.

Table 8 shows the measurements made with the test plants.

The levels of CI- concentration on a dry weight basis from the

rocket tests measured with a specific-ion electrode are sus­

piciously high. The last test (number 4) utilized a colorometric

method for Cl- analysis and is considered more reliable, but

appears also high. The uncertainty in the colorometric measure­

ments was approximately 15%. Applying the 15% uncertainty to the

last test, observe that the magnitude of the errors overlap, thus

rendering the net Cl-uptake data unreliable. Note, however, that

for all the test data, the trend of increased C1 concentration

is not violated. Therefore, the tests can be used to establish

overall trends.

In examining the data from corn, a direct correlation was

found between plant age and Cf uptake: the younger the plant,

the higher was the Cl- concentration on a dry weight basis. This

is consistent for each of the first three tests. With soybeans,

no correlation with age was observed. The net average HCI uptake

for corn was approximately 6000 pg/g dry weight for soybeans,

approximately 1000 pg/g. HC1 uptake in soybeans was reduced by

a factor of approximately six as compared to the monocotyledonous

plant represented by corn. These results are considered tenta­

tive because of the limited number of tests petformed.

The determination of real-time growth response to rocket

exhaust exposure is a first step in examining the more complex

plant responses which include among others reproduction and yield.

IIT RESEARCH INSTITUTE

58

Table 8

SUMMARY OF TEST PLANTS Hi UPTAKE MEASUREMENTS

Experimental

Rocket Number

Plant Type

Age* (days)

Exposure Time (min)

Rel. Hum. ()

Chamber HC1 Concentration

(ppm)

Dry Weight (g)

Total Cl Ng)

Total Cl-/ Dry Wt. (pg/g)

Net Cl-/ Dry Wt. (vg/g) Comments

1 beans 14 20 50 8 3.404 51,320 15,100 1,600 beans 14 control 2.113 28,560 13,500 corn 14 20 50 8 2.690 101,160 37,600 17,000 corn 14 control 4.273 87,960 20,600

2 beans 24 20 90 10 2.069 20,880 10,000 2,000 leaves pointed downward approximately 450

beans 24 control 2.877 23,040 8,000 corn 24 20 90 10 3.952 108,100 27,350 1,750 corn 24 control 3.254 83,440 25,600

3 beans 14 20 80 10 0.9019 6,850 7,600 910 leaves pointed downward approximately 45" '

beans 14 control 0.8346 5,580 6,690 corn corn'

14 14

20 control

80 10 1.2984 0.8656

19,200 8,770

14,800 10,100

4,700 wrinkles on leaf edges

4 beans 21 20 75 2 2.5 14,300 5,700 800 leaves pointed downward

beans 21 control 3.2 15,700 4,900 during exposure

corn 21 20 2 4.4 105,200 23,900 1,800 brown spots increased

corn 21 control 3.4 75,100 22,100 in area of leaf tips

*Measured from time of planting.

To gain some insight in this area, a preliminary growth response

experiment was conducted with soybeans in which internode elong­

ation was measured. Two sets of plants, the controls and plants

to be exposed were placed in corresponding chambers, the experi­

mental chamber and a temporary chamber, where identical light and

temperature conditions were maintained. Sensitive displacement

transducers were placed into the chambers and internode elongation

of two single plants was continuously recorded, before, during and

after the exposure to the rocket exhaust.

Figure 21 shows the results of the growth response measure­

ment. Immediately before and during the exposure to the rocket

exhaust, the elongation rate decreased to approximately 0.6 of

the overall elongation rate of the experimental plant. At the

midpoint of the 20 minute exposure, the elongation rate dropped

to 1/3 of this rate and approached slowly a zero rate. While it

is tempting to ascribe this decrease in the elongation rate to

the rocket exhaust, further tests are required.

The elongation rate of the control was approximately two

times higher than that of the experimental plant throughout the

test period. While such variations in elongation rates are not

uncommon, it excludes further interpretation of the results with­

out additional test. The manifestation of the growth response

in other plants after HCI exposures also remains to be investigated.

4.4 Metallic Corrosion Tests

Differential weight determinations were used to ascertain

the advent of significant corrosion. Table 9 lists all the data

obtained throughout the entire test program including visual

-observations. All the coupons received a single exposure to the

rocket exhaust effluent. To readily compare the data, Figure 22

was prepared.

In discerning trends from the differential weight data, note

first the general lack of positive increase. Usually when sig­

nificant corrosion occurs, weight gains are measured. Since this

lIT RESEARCH INSTITUTE

60

Note: Cause of growth reduction of the experimental

plant between 2:30 and 3:15 is unknown. Growth reduction starting at 3:42 may have resulted from exposure to rocket exhaust.

Plant

xposed

150

0

oIloo Experimental Plant

'4 Control Plan

05

0

12:30

GROW4TH

1:30

RESPONSE OF SOYBEAN

2:30

Time

Figure 21

SEEDLINGS TWO WEEKS AFTER

3:30

GERMINATION

4:30

0.04

0.03

0.02

Omild steel 0304 SS

0304 SS, space paint

Amild steel, lacquer

Nmild steel, enamel

02024 Al

66160 Al

OCommercial Al siding

open symbols, rocket i1 closed symbols, rocket #2

closed symbols, rocket #3

Environmental chamber conditions: T

RH = =

950F %100%

S to

0.01

S0.0

0 -0.01

-0.04 4-4i

-0.03 AA A

A

-0.04

0

II I

30 60 90

Exposure Time InsideEnvironmental Chamber (days)

Figure 22

-

120

NET COUPON WEIGHT CHANGES VERSUS LENGTH O'F EXPOSURE TIME

Table 9

MEASURED WEIGHT CHANGES ASSOCIATED WITH EXPOSED METAL COUPONS

Rocket Number

Sample Type*

Initial Coupon Weight

(gm)

Coupon Wt. Change After Exposure

(gin)

Coupon Wt. Change After

30 Days (gin)

Coupon Wt. Change After

60 Days (g)

Coupon Wt. Change After

120 Days .(gin)

Exposure Conditions

HCI Cone. %RH (ppm) Comments

1 mild steel 54.6085 +0.0012 0.0000 +0.0010 +0.0025 %50 8 rusted in environmental chamber 304 SS 304 SS, space paint mild steel lacquer mild steel enamel 2024 aluminum 6160 aluminum commercial Al siding

70.1722 72.9200 55.5674 57.4726 22.0987 21.2319 3.4024

-0.0007 -0.0053 -0.0112 -0.0056 +0.0020 -0.0002 -0.0004

-0.0031 -0.0066 -0.0211 -0.0171 +0.0027 -0.0006 -0.0001

-0.0036 -0.0081 -0.0249 -0.0234 +0.0013 -0.0017 -0.0016

-0.0019 -0.0068 -0.0260 -0.0236 +0.0029 +0.0006 +0.0001

'50 '50 N50 '50 %50 '50 N50

8 8 8 8 8 8 8

coating thickness = 8 mils

w

2 mild steel 304 SS 304 SS, space paint mild steel, lacquer mild steel, enamel 2024 aluminum 6160 aluminum commercial Al siding

53.8593 69.8119 72.9200 54.0404 56.8873 22.2523 20.2721 3.4697

+0.0014 -0.0021 -0.0036 -0.0195 -0.0052 +0.0003 -0.0004 -0.0006

+0.0037 -0.0040 -0.0056 -0.0294 -0.0189 +0.0007 -0.0004 -0.0007

+0.0038 -0.0029 -0.0057 -0.0311 -0.0201 -0.0004 -0.0013 -0.0015

+0.0056 -0.0029 -0.0052 -0.0333 -0.0255 -0.0014 +0.0006 +0.0012

'90 90 90 '0 '90 '90 '90 90

10 10 10 10 10 10 10 10

heavy rust in environmental chamber

coating thickness = 8 mils paint chipped during test

3 mild steel 304 SS 304 SS, space paint mild steel, lacquer mild steel, enamel 2024 aluminum 6160 aluminum

53.2960 69.1093 71.7960 54.4606 56.3894 22.1002 20.5965

+0.0029 -0.0014 -0.0099 -0.0140 -0.0039 +0.0029 +0.0003

+0.0024 -0.0033 -0.0108 -0.0176 -0.0097 +0.0024 -0.0001

+0.0029 -0.0034 -0.0107 -0.0222 -0.0138 +0.0021 -0.0008

+0.0038 -0.0038 -0.0107 -0.0286 -0.0269 +0.0027 -0.0002

80 "0 '90 '80 '80 '80 '80

10 10 10 10 10 10 10

rusted in environmental chamber

coating thickness = 8 mils

0

commercial Al siding 3.5465 -0.0004 -0.0004 -0.0001 -0.0002 80 10

4 mild steel 56.0566 +0.0003 -0.0010 +0.0004 --- %75 2 slight rusting after 30 days

304 SS 304 SS, space paint mild steel, lacquer mild steel, enamel 2024 aluminum

68.8626 73.6475 54.9344 57.8603 21.3193

-0.0019 -0.0055 -0.0184 -0.0249 -0.0010

-0.0034 -0.0076 -0.0257 -0.0249 -0.0012

-0.0024 -0.0075 -0.0266 -0.0321 -0.0005

---------------

'75 '75 75 75

N75

2 2 2 2 2

coating thickness = 8 mils

6160 aluminum commercial Al siding

20.1279 3.5052

-0.0005 -0.0001

-0.0007 -0.0003

-0.0006 0.0000

------

'75 '75

2 2

*Table 5 gives details regarding coupon metallurgical data preparation

did not occur with the exposed metallic coupons, significant cor­

rosion did not occur during the 120 day incubation period. Im­

portantly, the aluminum siding remained very constant, showing

no signs of corrosion. In contrast, several coupons lost a dis­

cernabie weight--both the painted mild steel coupons--but showed

no rust as did the unpainted mild steel. However, all the weight

differentials are small and even the trends cited must be considered

tentative.

4.5 Characterization and Scavenging of Rocket Exhaust

Aluminum Oxide Dust

Three instruments were used to obtain the particle size data:

the electrical aerosol analyzer (Thermo-Systems Modd1 3030), a

light-scattering instrument (Royco 220 and 245), and a cascade

impactor (Andersen). Each instrument is operable over the dif­

ferent particle diameter size ranges as indicated:

electrical aerosol analyzer, 0.01 to 0.7 pm

light scattering, 0.5 to 5.0 pm

cascade impactor, 1.0 to 9.0 Vm

Particle size differentiation with the EAA and the light­

scattering instruments was based on the geometric size of the

particle while size differentiation with the cascade' impactors

was based on particle inertia.

During the first scavenging test (Rocket #1) a light-scat­

tering instrument was employed (Royco 220). Use of this partic­

ular instrument was restricted by its low particle loading cap­

ability--the maximum particles concentration that can be accomo­.dated is 5 particles/cm3 However, the particle size pertaining

to the maixmum number concentration can be approximated and is

given below.

Time After Approx. Particle Rocket Ignition Size at Max Concentration

(min) Concentration (im) of HCl (ppm)

15 2.5 8.0

90 2.0 3.0

120 1.7 1.2

IIT RESEARCH INSTITUTE

64

Data for the second scavenging test were taken during the

third rocket firing. Figure 23 shows the data (corrected for

background) as obtained by the EAA and the light-scattering in­

strument (Royco Model 245). This light-scattering instrument had

adequate particle loading capability for the rocket exhaust cloud

being tested.

As can be seen in Figure 23, a substantial number of parti­

cles occur within the submicron particle size range. Note that

an inflection point occurs at approximately 0.7 pm, thus suggest­

ing a bimodal particle size distribution.' This is reasonable as

the complicated combustion process within the rocket nozzle gen­

erates particles by both gaseous condensation of alumina and by

the dispersion of molten alumina. In addition, Dr. Varsi's data

have been obtained through Ron Dawbarn at ARO, Inc. where a Titan

III exhaust cloud was sampled with a specially instrumented air­

craft (including an EAA) approximately 10 minutes after launch.

Varsi shows an inflection point at 0.5 tm--very comparable to our

own 0.7 pm inflection point. The total particle number concentra­

tion of the sampled Titan III exhaust cloud was 3 x 104 particles/3

cm whereas within our experimental chamber at 100 minutes after

ignition th concentration was 1.4 x 105 particles/cm3 and at 280

minutes, 1.2 x 105 particles/cm3 . The number concentrations dif­

fer by a factor of four. Not known is the HC concentration within

the exhaust cloud of the Titan III; this information can be used

to check the "dilution factor" of four. Based on the characteri­

zation of the Al203 particles, the exhaust from the JPL scale roc­

ket is considered as an accurate replica of the cloud generated

by a full-sized rocket. The reduction in particle count with time

is best accounted for by-coagulation of the submicron particles as

observed in Figure 23.

lIT RESEARCH INSTITUTE

65

io6

105_

104 -

4j

-o103

" 1 0 2 cJJ

0

101-0

101

1

approx. +L00 rain

[] approx. +280 min.

0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0

D (pm)

Figure 23 A1203 PARTICLE SIZE DATA FOR SECOND SCAVENGING TEST

Data from the Andersen impactor (flow rate = 28.2 1pm) are

given below.

Particle Size Range Mass Stage Number Collected (Nm) Collected (mg)

1 >9 0.8 2 5.5 - 9 0.2 3 3.3 - 5.5 0.1 4 2-3.3 0.2 5 1-2 1.0 6 <1 0.5

Note that, based on particle number the mean size is again in

the range of I to 2 pm. This collaborates the data obtained by

the other two techniques.

Measurements of particle size were also performed on the third rocket firing. The Royco Model 245 and the Andersen impactor were

used. The data from the Royco are shown in Figure 24 and the

Andersen data given below.

Particle Size Range Mass Stage Number Collected (pm) Collected (mg)

1 >9 0.0 2 5.5 - 9 0.0 3 3.3 - 5.5 0.1 4 2 - 3.3 0.1 5 1-2 0.3 6 <1 0.0

The particle size data obtained support the data obtained during

earlier scavenging tests.

During the last scavenging experiments, a sample of collected

rain drops, 1.1 mm indiameter, was set aside for analysis of the

entrained AI203 particles. Figure 25 displays the two particle

number distributions--both airborne A12 03 particles at the time

of the experiment and the AI2 03 particles in the rainwater. The

plot ifidicates that the droplets scavenge a larger proportion of

the smaller submicron particles than the larger particles.

The number concentration of Al2 03 particles during the scav­

enging experiment was 5.4 x 104 particles/cm as measured by the

light-scattering technique. The particle concentration within

the collected rain was 3.5 x 105 particles/cm3--approximately an lIT RESEARCH INSTITUTE

67

106

105

104

103

4jl

P,

S102

o + 30 min 6 ppm HCl RH = 74%.

o '+118 min 3 ppm HCI RH = 83%

Q +138 min 1.6 ppm HC1 RH = 83% 10 A +249 min 2.8 ppm HCL RH f 60%

0 I I

0.1 0.2 0.5 1.0 2.0 5.0 10.0 20.0 D (pm)

Figure 24 A1203 PARTICLE SIZE DATA FOR THIRD SCAVENGING TEST

60

50

-40 34

A Airborne Al 203

rA2 3

30 E Collected Rain A1203 30

Cd

0

0 0.2 0.5 1.0 2.0 10.0

Particle Size (pm)

Figure 25

A1 2 03 PARTICLE SIZE DISTRIBUTIONS BASED ON NUMBER FOR BOTH AIRBORNE AND COLLECTED RAIN (1.1 mm DIAMETER)

order-6f-magnitude greater than the airborne concentrationl. This

result appears questionable as the volume any one dVop sweeps3. For the rolet to ql­

through during the fall is only 1.24 cm

lect the measured number of particles, the effective volume swept

must be about ten times as great. A tentative explanation i, the

AI203 aerosol is electrically charged--as are most combustion

aerosols--in such a way as to be attracted to the falling droplets.

[Another plausible explanation may be added collection due to the

eddy currents created by thq falling droplets.J This experiment

was conducted only once to demqnstrate A1203 scavenging. Further

investigation of Al203 sdavenging would require significant improve­

ment of the experimental procedure used here while maintaining the

use of actual rocket exhaust to conduct thp experiment.

Figures 26 and 27 are photomicrographs from a scanning electron

microscope showing the Al203 particles collected by the raindrops.

The non-spherical particles are not AI203 particles, but filter

debris resulting from sample preparation. The crystalline mater­

ial adjacent to a number of the AI203 particles contains signifi$

cant quantities of chlorine (as determined by SEM microprobe) and

appears to be a chlcrine-containin$ salt. Othqr partiples not

possessing adjacent crystals may not be properly oriented, or,

indeed, may not have adjoining chlorine crystals. An explqnation

for the adjacent crystals is not apparent.

5. CONCLUSION

Rocket exhaust HCl scavenging tests, surface corros$on, and

uptake studips were undertaken experimentally. The conclusions

presented here are developed and discussed within thp text of this

report. The conclusions are as follows:

1. The washout coefficient for HC1 scavenging determined empirically is:

5) MA0 .176 R0.773A = (5.12 x 10-

The nonlinearity effect of the M Ol7 factor is weak, varying only a factor of two ove% the range of HCI concentrations tested--0.2 to 10.0 ppm. This factor of two is overshadowed by the uncertainty comprising the experimental data.

IIT RE5EARq1 INSTI'TUTE

70

ORIGINAL PAGE 1b OF POOR QUALITY

Figure 26

PHOTOMICROJGRAPHS.,SHOWING A1203 PARTICLES COLLECTED BY

1.1Im RAINDROPS AT 3,000X MAGNIFICATION

Figure 27

PHOTOMICROGRAPHS SHOWING A1203 PARTICLES COLLECTED BY

1.1i mm RAINDROPS AT 10,000 MAGNIFICATION IIT RESEARCH INSTITUTE

71

2. The surface uptake (calculated deposition velocities) rates for distilled water is less than the ieference 0.05 normal NaOH solution and concentration dependent. Measurement of sea water and brackish water HCI depo­sition velocity is untenable due to inherently high chlorine background vAlues.

3. Plant HC1 uptake measurements utilizing corn and soy­beans indicate significant uptake trends. Plant age effectively correlates the corn data: younger the plant age, the higher the Cl- concentration on a dry weight basis. (There was no apparent correlation 6f plant age and HC1 uptake for the soybeans.) Real-time growth response of soybeans occurs with HC1 exposure of 20 minutes duration.

4. No significant corrosion is apparent from the metallic coupon corrosion experiments.

5. Particle size characteristics of Al 0 small-scale rocket (227 g) is comparable to partile size data from full size rockets. It was tentatively found that raindrops effectively scavenge AI203 particulate mater­ial.

lIT RESEARCH INSTITUTE

72

Appendix A

PURE GASEOUS HC1 SCAVENGING

IIT RESEARCH INSTITUTE

73

PURE GAS HCl SCAVENGING

At the conclusion of the last rocket exhaust scavenging test,

pure HC1 was introduced to the experimental chamber via lecture

bottles pressurized with HC. Three rain drop sizes were tested-­

0.55 mm, 1.1 mm, and 3.0 mm--with the results shown in Figure A-I.

Immediately apparent is the lack of agreement with the empirical

correlation resultant from the rocket exhaust scavenging data (an

order-of-magnitude below the pure gaseous H1 scavenging data).

Experiments with actual rocket exhaust constituents are, therefore,

seen as highly significant.

The experimental conditions consisted of high relative humidi­

ties, in excess of 95%. And, indeed, here an extremely visible

and persistent acid-mist was formed with the introduction of pure

HC1 into the experimental chamber. This acid-mist was. not visible

or detected with the instrumentation under rocket exhaust HCl.

Apparently, the interacting constituents of the rocket exhaust

cloud serve to suppress the scavenging rate of HCI for a given

HCI concentration.

1IT RESEARCH INSTITUTE

74

ixl0 - 4

A BA

Ix10- 5

El

o ix10-6

A

sd

(672 x 10-5 0824

ix10- 7 ,D

Raindrop Size

SAD-- 0.55 mm

=1.1 mm

D = 3.0 mm

Ix 0 ­

0.1 0.2

I I I I

0.5 1.0 2.0 5.0 10.0 20.0 50.0 100

Corrected HCI Concentration (ppm)

Figure A-i

PURE HCI SCAVENGING DATA FOR MULTIPLE RAINDROP SIZES

I

200 500 1000

REFERENCES

1. Pellett, G.L. "Washout of HC1 and Application to Solid Rocket Exhaust Clouds", paper presented at the Precipitation Sdaveng­ing Symposium, University of Illinois at Urbana, Illinois (October 14-18, 1974).

2. Knutson, E.O.; Fenton, D.L., "Atmospheric Scavenging of Hydro­chloric Acid, "NASA CR-2598, George C. Marshall Space Flight Center, Alabama, IIT Research Inst., Chicago, Ill. (August 1975).

3. Dumbauld, R.K.; Bjorkland, J.R.; Bowers J.F.; "NASA/MSFC Mul­tilayer Diffusion Models and Computer Program for Operational Prediction of Toxic Fuel Hazards", H.E. Cramer Co., Report No. TR-73-301-02 to the Gqorge C. Marshall Flight Center, (March 1973).

4. Ranade, M.B., "Sampling Interface for Quantitative Transport of Aerosols", Final Report, IITRI, EPA Contract No. 68-02-0579, Chem. and Physics Lab., EPA, Research Triangle Park, N.C.

5. Strand, Leon, Private Communication, Jet Propulsion Labora­tory, Pasedena, California, June 9, 1977.

6. Lord Rayleigh, Philosoph. Mag., 36 (1878).

7. Strom, L., Rev. Sci. Instruments, 35: 778 (1969).

8. Berglund, R.N.; Liu, B.Y.H., Environmental Sci. Technol., 7: 747 (1973).

9. Schneider, J.M.; Hendricks, C.D., Rev. Sci. Instruments, 35: 1349 (1964)

10. Wedding, J.B.; Stukel, J.J., "Operational Limits of Vibrating Orifice Aerosol Generator", Environmental Sci. Technol., 8 (5): 456 (1974).

11. Gunn, R.; Kinzer, G.C., "The Theoretical Velocity of Fall for Water Droplets in Stagnant Air", J. Meteor, 6: 243-248 (1949).

12:. Knutson, E.O.; Fenton, D.L., "Atmospheric Scavenging of Hydro­chloric Acid", NASA CR-2598, George C. Marshall Space Flight Center, Alabama, IIT Research Inst., Chicago, Ill. (August 1975).

13. Dawbarn, R., Private Communication, ARO Inc., Arnold Air Force Station, Tenn., May 18, 1976.

14. Sherwood, T.K., Pigford, R.L., and Wike, C.R., Mass Transfer, McGraw-Hill, USA, pp. 81-83 (1975).

15. Fuchs, N.A., The Mechanics of Aerosols, Pergamon Press, Oxford

pp. 184, 204-206 (1964)

jit RESEARCH INSTITUTE

76

REFERENCE (con.)

16. Gillespie, G.R;; Johnstone, HlF.,- "Particle Size Distribu­tion in Some Hygroscopic Aerosols", Chem. Engr. Prog., pp. 74F-80F, Feb. 1955.

17. Fenton, D.L.; Ranade M.B., "Aerosol Formation Threshold for the HCi-Water Vapor System", Environmental Si. and Teqhnol., 10 (12): 1160-1162 (Nov. 1976).

18. Schem, C., Private Communication, Naval Air Propulsion Test Center, Trenton, New Jersey (October 19, 1976).

19. Marshall, J.S.; Palmer, W.M., "The Distribution of Raindrops with Size", J. Meteor., 5: 165-166 (1948).

20. Dingle, A.N.; Hardy, K.R., "The Description of Rain by Means of Sequential Raindrop Size Distribution", Quart. J. Roy. Meteor. Soc., 88: 301-304 (1962).

21. Buchanan, P., -M.D., presentation at "Environmental Effects Workshop", Atmospheric Diffustion/Environmental Effects Technical Task Team, ES43, NASA/Marshall Space Flight Center, Alabama (September 8, 1974).

22. Heck, W., private communication, North Carolina State Univers­ity (September 8, 1976).

23. Lerman, S.; Taylor, O.C.; Darley, E.F., "Phytototoxicity of Hydrogen Chloride Gas with a short~term exposure", Atmospheric Environ., 10: 873-878 (1976).

IIT RESEARCH INSTITUTE

77

NASA FORMAL REPORT

FFNo 665 Aug 65